Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Amos, At 23.35 02/10/2007, Amos Jeffries wrote: BTW, Which OS was this on? That particular bug props up when some IPv6 extensions to local resolver are not available for use. Windows, using MSYS+MinGW. I meant: 2K/NT, XP, Vista? I'd kind of expect this on 2kK/NT, possibly XP as its IPv6-related. But the MinG for Vista should be emulating a v6 capable resolver. Sorry for the delay in the response ... :-( Let me to clarify something about the Squid Windows support. Currently Squid can be compiled in Windows using 3 different ways: - Cygwin - MSYS+MinGW - MS Visual Studio (version 6 for Squid 2.6, version 2005 for Squid 3) Cygwin and MSYS+MinGW are officially supported on the main distribution of 2.6 and 3.0. Both are a bash+gcc based build environment, so the build process is identical to all other platforms. The MS Visual Studio build environment is supported from the SQUID_NT_2_6 stable branch on squid-cache.org and from the nt branches on Sourceforge. There is an essential difference between Cygwin and MSYS+MinGW: - Cygwin is based on a run-time emulation layer, so using Cygwin we have a better source compatibility with worse performance and Windows integration and to run, the executables need the cygwin run-time called cygwin1.dll. - MSYS+MinGW is based on a sort of build-time emulation, so there is a worse compatibility, but the performance are very good. At run-time, the executables need only the standard Windows C Runtime Library, msvcrt.dll, available on any Windows machine. MinGW doesn't provide any kind of emulation layer. So when speaking about Windows, the used operating system is significative only at run-time, not at build-time. Using MSYS+MinGW and MS Visual Studio, the Microsoft IPv6 support must be enabled raising the target platform from standard win32 to Windows XP/2003, but the resulting executable cannot run on previous Windows versions. For the IPv4 only build test, the win32 standard target must be used. Regards Guido - Guido Serassio Acme Consulting S.r.l. - Microsoft Certified Partner Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY Tel. : +39.011.9530135 Fax. : +39.011.9781115 Email: [EMAIL PROTECTED] WWW: http://www.acmeconsulting.it/
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Amos, At 12.46 01/10/2007, Guido Serassio wrote: Guido, that was a feature. :-p It was meant to provide an autoconf arranged function-macro 'ABC(i)' that allows the nsaddrs to be indexed like an array instead of conditionally coded for the various access methods. There was a regression bug in *nix autoconf 2.59 that prevents it working. But that was fixed again by 2.61+ $ autoconf -V autoconf (GNU Autoconf) 2.59 Written by David J. MacKenzie and Akim Demaille. Arg !!! I went crazy for 4 hours trying to understand a not-sense C++ build error . :-( If there is any way to get it to work that would be good. Otherwise there are quite a number of lines to be added to dnsserver.cc to cope with the different varitations. The next planned version of autotools on MinGW are: - autoconf 2.61 - automake 1.10 I will try with the preview builds to see what happens. Just upgraded autoconf and automake and now my development environment is broken: I cannot no more bootstrap squid . :-( I think that we must find another solution. I've moved the glue from autoconf into the source. It should work with autoconf 2.59 now. Amos
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Amos, At 23.14 01/10/2007, Amos Jeffries wrote: Arg !!! I went crazy for 4 hours trying to understand a not-sense C++ build error . :-( Only 4 hours? man your lucky. It took G. Panula and I a nasty few weeks to figure this one out. The error happens in MemPool.cc, so I have compared all the involved files against a working squid 3, but i took 4 hours before comparing autoconf.h :-) BTW, Which OS was this on? That particular bug props up when some IPv6 extensions to local resolver are not available for use. Windows, using MSYS+MinGW. If there is any way to get it to work that would be good. Otherwise there are quite a number of lines to be added to dnsserver.cc to cope with the different varitations. The next planned version of autotools on MinGW are: - autoconf 2.61 - automake 1.10 FWIW, I have been using 2.61 here no problems. Panula has yet to get back to me on if his/her upgrade to 2.61 has resulted in a working autoconf.h . The bypass for now is to configure, then patch the autoconf.h changing /* #undef _SQUID_RES_NSADDR6_LIST */(i) to: /* #undef _SQUID_RES_NSADDR6_LIST(i) */ I will try with the preview builds to see what happens. If that fails, I'll work up some glue. What about fix this using sed into bootstrap.sh ? Regards Guido - Guido Serassio Acme Consulting S.r.l. - Microsoft Certified Partner Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY Tel. : +39.011.9530135 Fax. : +39.011.9781115 Email: [EMAIL PROTECTED] WWW: http://www.acmeconsulting.it/
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Henrik, At 23.29 01/10/2007, Henrik Nordstrom wrote: On mån, 2007-10-01 at 20:53 +0200, Guido Serassio wrote: Just upgraded autoconf and automake and now my development environment is broken: I cannot no more bootstrap squid . :-( What happens? Broken packages . Just reverted to release (and working ) versions Regards Guido - Guido Serassio Acme Consulting S.r.l. - Microsoft Certified Partner Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY Tel. : +39.011.9530135 Fax. : +39.011.9781115 Email: [EMAIL PROTECTED] WWW: http://www.acmeconsulting.it/
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Amos, At 22.03 30/09/2007, Amos Jeffries wrote: Update of cvs.devel.squid-cache.org:/cvsroot/squid/squid3 Modified Files: Tag: squid3-ipv6 configure.in Log Message: Fix syntax error in nsaddr tests Guido, that was a feature. :-p It was meant to provide an autoconf arranged function-macro 'ABC(i)' that allows the nsaddrs to be indexed like an array instead of conditionally coded for the various access methods. There was a regression bug in *nix autoconf 2.59 that prevents it working. But that was fixed again by 2.61+ $ autoconf -V autoconf (GNU Autoconf) 2.59 Written by David J. MacKenzie and Akim Demaille. Arg !!! I went crazy for 4 hours trying to understand a not-sense C++ build error . :-( If there is any way to get it to work that would be good. Otherwise there are quite a number of lines to be added to dnsserver.cc to cope with the different varitations. The next planned version of autotools on MinGW are: - autoconf 2.61 - automake 1.10 I will try with the preview builds to see what happens. Regards Guido - Guido Serassio Acme Consulting S.r.l. - Microsoft Certified Partner Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY Tel. : +39.011.9530135 Fax. : +39.011.9781115 Email: [EMAIL PROTECTED] WWW: http://www.acmeconsulting.it/
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
On mån, 2007-10-01 at 20:53 +0200, Guido Serassio wrote: Just upgraded autoconf and automake and now my development environment is broken: I cannot no more bootstrap squid . :-( What happens? Regards Henrik signature.asc Description: This is a digitally signed message part
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Hi Amos, At 12.46 01/10/2007, Guido Serassio wrote: Guido, that was a feature. :-p It was meant to provide an autoconf arranged function-macro 'ABC(i)' that allows the nsaddrs to be indexed like an array instead of conditionally coded for the various access methods. There was a regression bug in *nix autoconf 2.59 that prevents it working. But that was fixed again by 2.61+ $ autoconf -V autoconf (GNU Autoconf) 2.59 Written by David J. MacKenzie and Akim Demaille. Arg !!! I went crazy for 4 hours trying to understand a not-sense C++ build error . :-( If there is any way to get it to work that would be good. Otherwise there are quite a number of lines to be added to dnsserver.cc to cope with the different varitations. The next planned version of autotools on MinGW are: - autoconf 2.61 - automake 1.10 I will try with the preview builds to see what happens. Just upgraded autoconf and automake and now my development environment is broken: I cannot no more bootstrap squid . :-( I think that we must find another solution. Okay, I have an idea for some glue. I'll get right on it. Amos
Re: squid3-ipv6 squid3 configure.in,1.63.2.61,1.63.2.62
Update of cvs.devel.squid-cache.org:/cvsroot/squid/squid3 Modified Files: Tag: squid3-ipv6 configure.in Log Message: Fix syntax error in nsaddr tests Guido, that was a feature. :-p It was meant to provide an autoconf arranged function-macro 'ABC(i)' that allows the nsaddrs to be indexed like an array instead of conditionally coded for the various access methods. There was a regression bug in *nix autoconf 2.59 that prevents it working. But that was fixed again by 2.61+ If there is any way to get it to work that would be good. Otherwise there are quite a number of lines to be added to dnsserver.cc to cope with the different varitations. Amos Index: configure.in === RCS file: /cvsroot/squid/squid3/configure.in,v retrieving revision 1.63.2.61 retrieving revision 1.63.2.62 diff -C2 -d -r1.63.2.61 -r1.63.2.62 *** configure.in 26 Sep 2007 03:13:26 - 1.63.2.61 --- configure.in 30 Sep 2007 18:16:13 - 1.63.2.62 *** *** 3278,3282 ac_cv_have_res_ext_nsaddr_list=no)) if test $ac_cv_have_res_ext_nsaddr_list = yes ; then ! AC_DEFINE([_SQUID_RES_NSADDR6_LIST(i)],[_res_ext.nsaddr_list[[(i)]].sin6_addr],[If _res_ext structure has nsaddr_list member]) AC_DEFINE(_SQUID_RES_NSADDR6_COUNT,ns6count,[Nameserver Counter for IPv6 _res_ext]) fi --- 3278,3282 ac_cv_have_res_ext_nsaddr_list=no)) if test $ac_cv_have_res_ext_nsaddr_list = yes ; then ! AC_DEFINE([_SQUID_RES_NSADDR6_LIST],[_res_ext.nsaddr_list[[(i)]].sin6_addr],[If _res_ext structure has nsaddr_list member]) AC_DEFINE(_SQUID_RES_NSADDR6_COUNT,ns6count,[Nameserver Counter for IPv6 _res_ext]) fi *** *** 3305,3309 ac_cv_have_res_ext_nsaddrs=no)) if test $ac_cv_have_res_ext_nsaddrs = yes ; then ! AC_DEFINE([_SQUID_RES_NSADDR6_LIST(i)],[_res._u._ext.nsaddrs[[(i)]]-sin6_addr],[If _res structure has _ext.nsaddrs member]) AC_DEFINE(_SQUID_RES_NSADDR6_COUNT,[_res._u._ext.nscount6],[Nameserver Counter for IPv6 _res]) fi --- 3305,3309 ac_cv_have_res_ext_nsaddrs=no)) if test $ac_cv_have_res_ext_nsaddrs = yes ; then ! AC_DEFINE([_SQUID_RES_NSADDR6_LIST],[_res._u._ext.nsaddrs[[(i)]]-sin6_addr],[If _res structure has _ext.nsaddrs member]) AC_DEFINE(_SQUID_RES_NSADDR6_COUNT,[_res._u._ext.nscount6],[Nameserver Counter for IPv6 _res]) fi
